energies
Article Experimental Study and Conjugate Heat Transfer Simulation of Pulsating Flow in Straight and 90◦ Curved Square Pipes
Guanming Guo 1 , Masaya Kamigaki 1, Yuuya Inoue 1, Keiya Nishida 2, Hitoshi Hongou 3, Masanobu Koutoku 3, Ryo Yamamoto 3, Hideaki Yokohata 3, Shinji Sumi 3 and Yoichi Ogata 2,*
1 Graduate School of Engineering, Hiroshima University, Hiroshima 7390046, Japan; [email protected] (G.G.); [email protected] (M.K.); [email protected] (Y.I.) 2 Graduate School of Advanced Science and Engineering, Hiroshima University, Hiroshima 7390046, Japan; [email protected] 3 Mazda Motor Corporation, Hiroshima 7308670, Japan; [email protected] (H.H.); [email protected] (M.K.); [email protected] (R.Y.); [email protected] (H.Y.); [email protected] (S.S.) * Correspondence: [email protected]
Abstract: The turbulent pulsating flow and heat transfer in straight and 90◦ curved square pipes are investigated in this study. Both experimental temperature field measurements at the cross-sections of the pipes and conjugate heat transfer (CHT) simulation were performed. The steady turbulent flow was investigated and compared to the pulsating flow under the same time-averaged Reynolds number. The time-averaged Reynolds number of the pulsating flow, as well as the steady flow, was approximately 60,000. The Womersley number of the pulsating flow was 43.1, corresponding to a 30 Hz pulsating frequency. Meanwhile, the Dean number in the curved pipe was approximately Citation: Guo, G.; Kamigaki, M.; 31,000. The results showed that the local heat flux of the pulsating flow was greater than that of the Inoue, Y.; Nishida, K.; Hongou, H.; steady flow when the location was closer to the upstream pulsation generator. However, the total heat Koutoku, M.; Yamamoto, R.; flux of the pulsating flow was less than that of the steady flow. Moreover, the instantaneous velocity Yokohata, H.; Sumi, S.; Ogata, Y. and temperature fields of the simulation were used to demonstrate the heat transfer mechanism of Experimental Study and Conjugate the pulsating flow. The behaviors, such as the obvious separation between the air and pipe wall, the Heat Transfer Simulation of Pulsating Flow in Straight and 90◦ Curved low-temperature core impingement, and the reverse flow, suppress the heat transfer. Square Pipes. Energies 2021, 14, 3953. https://doi.org/10.3390/en14133953 Keywords: pulsating flow; curved pipe; temperature fields; conjugate heat transfer; local heat transfer
Academic Editor: Gabriela Huminic
Received: 27 May 2021 1. Introduction Accepted: 28 June 2021 Fluid flow and heat transfer in pulsating flows occur in a variety of engineering Published: 1 July 2021 applications, such as reciprocating engines, electronics cooling, and certain heat exchangers. Arterial blood flow is a similar example of such flows. In recent decades, pulsating flow Publisher’s Note: MDPI stays neutral and its heat transfer have become the subjects of interest. with regard to jurisdictional claims in The flow in the intake and exhaust manifolds of an automotive intake and exhaust published maps and institutional affil- system is a typical pulsating flow. The flow and heat transfer characteristics of the intake iations. and exhaust manifolds directly affect the efficiency of the downstream catalyst. Increasingly stringent emission standards require high efficiency of exhaust catalysts. According to Host et al. [1], management of exhaust system heat losses is critical for after-treatment system control and optimization. Therefore, the study of pulsating flows and their heat Copyright: © 2021 by the authors. transfer is a significant problem for optimizing the design of automotive exhaust systems. Licensee MDPI, Basel, Switzerland. Experimental, analytical, and numerical studies on fluid flow and heat transfer of This article is an open access article pulsating flows have been conducted by numerous researchers. The characteristics of distributed under the terms and laminar pulsating flow in a long pipe with constant wall heat flux were experimentally and conditions of the Creative Commons numerically investigated by Zhao and Cheng [2]. The results of the experiment and numer- Attribution (CC BY) license (https:// ical solution, such as the time-resolved fluid temperature and space-cycle averaged Nusselt creativecommons.org/licenses/by/ number, were similar. An experimental study was carried out for laminar pulsating pipe 4.0/).
Energies 2021, 14, 3953. https://doi.org/10.3390/en14133953 https://www.mdpi.com/journal/energies Energies 2021, 14, 3953 2 of 20
flow with constant wall heat flux by Habib [3]. It has been reported that, compared with the Reynolds number, the Nusselt number is strongly affected by the pulsation frequency. However, an analytical method for pulsating laminar convection heat transfer in a pipe was studied by Yu et al. [4]. The results showed that pulsation does not affect the time- averaged Nusselt number. Guo and Sung [5] studied pulsating laminar flows with different amplitudes. It was observed that for small amplitudes, the Nusselt numbers in various forms lead to inconsistent results; however, at higher amplitudes, heat transfer is always enhanced. Theory analysis studied by Hsiao [6] of a steady, two-dimensional, incompress- ible, micropolar, and nanofluid laminar flow reported distributions of the local convective heat transfer coefficient and the temperature of the stretching sheet. Many dimension- less parameters such as the Prandtl number, the Eckert number, the Brownian motion number et al. influence the heat and mass transfer performance of the stretching sheet. For turbulent pulsating flow, the characteristics of convection heat transfer in a pipe with a constant wall temperature were numerically studied by Wang and Zhang [7]. It has been reported that the Womersley number is a critical parameter in the study of pulsating flow and heat transfer. An experimental investigation of pulsating turbulent flow in a pipe with constant heat flux was studied by Elshafei et al. [8]. The results showed that the total Nusselt number was strongly affected by both the Reynolds number and pulsation frequency. The value of the local Nusselt number either increased or decreased over the steady flow value. The heat transfer in an oscillating flow in a cylindrical pipe was experimentally studied by Bouvier et al. [9]. It was observed that the determination of the heat flux by using wall temperatures exclusively or fluid temperatures measured at different radial positions yields different results. In addition to the conditions of different Reynolds numbers and pulsation frequencies, the influence of pulsator location on heat transfer in a pipe with constant heat flux was experimentally investigated by Patel and Attal [10]. The results indicated that the Nusselt number was strongly affected by both the pulsation frequency and Reynolds number when the pulsation mechanism was downstream of the tested pipe exit. The investigation by Wantha [11] also reported that both the pulsation frequency and amplitude of the pulsating flow play important roles in the heat transfer of pulsating flows. Nishandar et al. [12] highlighted that higher values of the local heat transfer coefficient occurred at the entrance of the tested pipe, and the mean heat transfer coefficient increased if pulsation was created in the flow of air. An experimental study of convective heat transfer and pressure drop in a helically coiled tube was investigated by Khosravi-Bizhaem et al. [13]. The findings indicated that the greatest heat transfer intensification occurred at a low Reynolds number, where the pulsating flow could affect the laminar boundary layer significantly compared with the turbulent pulsating flow. From these studies, it can be observed that many factors affect the heat transfer in pulsating flows. However, the main physical mechanisms of these factors’ influence on heat transfer have not been fully described. An experimental investigation by Shiibara et al. [14] applied a technique using infrared thermography to measure the spatio–temporal heat transfer to clarify the mechanism of heat transfer enhancement in a pipe flow upon sudden acceleration and deceleration. An experimental study by Plotnikov and Zhilkin [15] on the thermal and mechanical characteristics of the pulsating flow in the intake system of a turbocharged piston engine reported the influence of external turbulence on the thermal and mechanical characteristics of pulsating flows. Simonetti et al. [16] experimentally investigated the instantaneous characteristics of the flow field and heat transfer in pulsating flows representative of engine exhaust flow operating conditions. A 1D analytical formulation of convective heat transfers in engine exhaust-type turbulent pulsating flows was also presented. The velocity amplitude ratio was identified as representative of the predominant heat transfer enhancement mechanism. Although there have been many studies on heat transfer in pulsating flow, almost all of them are concentrated in straight pipes. Curved pipes as a popular “passive enhancement” technique for heat transfer are widely used in a wide range of industrial applications [17]. Energies 2021, 14, 3953 3 of 20
The steady turbulent flow and heat transfer in the curved pipe was studied by Guo et al. [18]. Under the same geometric shape of the curved pipe, the present study also conducted the experiment of steady flow. Compared with the reference [18], the length of the test section and the temperature measurement points in each cross-section were increased in the present study to get more detailed temperature field information. Many researchers have studied the instantaneous velocity field of pulsating flow in curved pipes [19–22]. The pulsating flow in curved pipes is more complicated than that in straight pipes. Unsteady secondary motions, such as Dean vortices and Lyne vortices, were observed in these studies. Few researchers have investigated the heat transfer characteristics of pulsating flows in curved pipes. The pulsating flow in the tailpipe of a valveless Helmholtz pulse combustor was investigated by Zhai et al. [23]. Based on experimental data, a Nusselt correlation was proposed for the pulsating flow in the straight, 45◦, 90◦, and 135◦ curved tailpipe. Using the same experimental instrument, Xu et al. [24] studied the pulsating flow in the curved tailpipe. This study also included a simulation verification. The flow-reversing, Dean vortex forming, shedding, and reforming processes at the inlet and exit of the curved section contribute to convective heat transfer enhancement. To conclude this review, it should be noted that the measurements of temperature fields inside pipes have rarely been reported owing to the difficulty of measuring the internal temperature of pipes in pulsating flow. In curved pipes, owing to the secondary flow, the internal temperature field inevitably changes after the curvature part, which leads to changes in the local heat transfer characteristics. The catalyst is installed downstream of the exhaust manifold; therefore, the study of the temperature field inside the straight and curved pipes can be used to improve the efficiency of the catalyst. In addition, in the existing studies, the constant wall heat flux thermal boundary condition was used by Zhao and Cheng [2], Habib [3], Yu et al. [4], Guo and Sung [5], Elshafei et al. [8], Patel and Attal [10], and Khosravi-Bizhaem et al. [13]. The constant wall temperature thermal boundary condition was used by Wang and Zhang [7] and Bouvier et al. [9]. Few researchers have studied the heat transfer of pulsating flow in curved pipes based on the third type of wall thermal boundary condition (the heat transfer coefficient of the wall and ambient temperature are defined). Automotive intake and exhaust manifolds are directly exposed to the environment. This situation is more appropriate for this type of wall thermal boundary condition. Therefore, the third type of wall thermal boundary condition was employed to the numerical simulation in this study. Both experiment and simulation have been applied to investigate the turbulent pulsat- ing flow and heat transfer in straight and 90◦ curved pipes, which are often used in engine manifolds. The aim of this study is to clarify the heat transfer performance of pulsating flow in straight and 90◦ curved pipes and analyze the heat transfer mechanism in combination with numerical simulation. The steady turbulent flow was used as a comparison to confirm whether the pulsating flow enhances or suppresses heat transfer. In the experiment, a hot air generator was applied to produce the high-temperature air of the inlet conditions. The pulsating flow was produced by a steady flow, together with a pulsation generator. The velocity and temperature fields were measured by hot-wire anemometers and thermocouples separately. The numerical simulation reported the instan- taneous temperature field and velocity field of the pulsating flow, as well as the effects of sec- ondary flow and flow impingement in the curved pipe on the heat transfer characteristics.
2. Experimental Method 2.1. Experimental Setup The experimental device diagram shown in Figure1 was designed and applied to investigate the influences of the pulsating flow in straight pipe and 90◦ curved pipes on forced-convection heat transfer. Air as an operating fluid is heated by a hot air generator (HAP 3100, Hakko Electric, Tokyo, Japan) in this open-loop system and passed through the pipes to the atmosphere. This experimental system can provide both steady and pulsating flows. As shown in the bottom left of Figure1, the pulsating flow was generated by rotating Energies 2021, 14, 3953 4 of 20
a disk, and four vent holes were arranged at even intervals on the disk. The frequency of the pulsating flow was realized by an AC motor that controlled the rotating speed of the disk. The steady flow was achieved by keeping the disk at a specific position where the hole on the disk was completely opened. The coordinate systems X, Y, and Z represent the streamwise direction, horizontal direction of the cross-section of the pipes, and vertical Energies 2021, 13, x FOR PEER REVIEW 5 of 22 direction of the cross-section of the pipes, respectively.
Figure 1. DiagDiagrammaticrammatic sketch sketch ofof experimental experimental apparatus apparatus and and coordinate coordinate system system..
T34Schematic K-type diagramsthermocouples of the were test used pipes to are measure shown the in Figurehot air 2time. The-averaged pipes were temper- made of aluminum,ature of each and cross the-section. processed As shown inner in and Figure outer 3c, diameters each section of comprised the pipes were36 temperature 32 and 40 mm, respectively.measurement points. According The straight to [18], and there curved are not pipes enough have 12 temperature and 15 temperature measuring measure- points in thement area cross formed-sections, by respectively. secondary flow T32 toK- capturetype thermocouples the detailed were temperature used to measure distribution. the In thehot presentair temperature study, theoscillations number under of measurement pulsating flow. holes was increased to 126. Among these holes, 120 holes with a diameter of 1.6 mm were drilled on the top surface of the pipes for the K-type thermocouple (T34, Okazaki, Kobe, Japan) insertion to measure the air temperature, and six holes with a diameter of 8 mm were drilled for another type of K-type thermocouple (T32, Okazaki, Kobe, Japan) insertion to measure the temperature oscillation under pulsating flow. As shown in Figure3b, the T32 K-type thermocouple is composed of a 13 µm thermocouple and a 25 µm thermocouple. Two holes with a diameter of 5 mm were drilled on the bottom surfaces of the upstream and downstream pipes, respectively. A hot-wire anemometer (0251R-T5, Kanomax, Andover, NJ, USA) was installed in these two holes to measure the inlet and outlet velocities of the hot air after calibration with a pitot tube under the high temperature. Figure3a,b shows the methods used to measure the velocity and temperature, respectively. The temperature field and velocity are measured separately as in Figure3a. The temperature field was measured after the velocity condition had been measured, so the influence of the size of the hot wire on the temperature field measurement of the test part can be ignored. The temperature fields of corresponding cross-sections as shown in Figure2 were measured one by one, and the temperatures of the 36 points in each cross-section as shown in Figure3c were measured point by point. Therefore, the influence of the size of the T34 K-type thermocouple on the measurement results can be ignored. Although the T32 K-type thermocouple as shown in Figure3b is relatively large, it will affect the downstream flow field and temperature distribution. Thus, we measured the temperatures of corresponding cross-sections one by one. When a hole in Figure3b is inserted by the thermocouple, the other holes are filled with screws to reduce Figure 2. Diagrammatic sketch of measurement points on pipes (top view): (a) straight pipe; (b) thecurved measurement pipe; (c) upstream error. pipe; (d) downstream pipe [18].
Energies 2021, 13, x FOR PEER REVIEW 5 of 22
Figure 1. Diagrammatic sketch of experimental apparatus and coordinate system.
T34 K-type thermocouples were used to measure the hot air time-averaged temper- ature of each cross-section. As shown in Figure 3c, each section comprised 36 temperature Energies 2021, 14, 3953 measurement points. The straight and curved pipes have 12 and 15 temperature measure- 5 of 20 ment cross-sections, respectively. T32 K-type thermocouples were used to measure the hot air temperature oscillations under pulsating flow.
Energies 2021, 13, x FOR PEER REVIEWFigureFigure 2. Diagrammatic 2. Diagrammatic sketch of sketch measurement of measurement points on pointspipes (top on pipesview): (top(a) straight view): pipe; (a) straight 6(b of) 22 pipe; curved pipe; (c) upstream pipe; (d) downstream pipe [18]. (b) curved pipe; (c) upstream pipe; (d) downstream pipe [18].
FigureFigure 3. 3. CrossCross-section-section of of pipes pipes (from (from upstream upstream view view):): (a (,ab,)b measurement) measurement methods; methods; (c) ( cmeasure-) measurement mentpoints points [18]. [18].
AsT34 shown K-type in thermocouplesFigure 2, to facilitate were the used subsequent to measure discussion, the hot airthe time-averaged first measurement tempera- sectionture of named each cross-section. “Inlet” at the Asoutlet shown of the in upstream Figure3c, pipe each is sectiondefined comprisedas the origin. 36 S1 temperature to S10 representmeasurement the ten points. measured The cross straight-sections and curvedin the straight pipeshave pipe. 12The and four 15 sections temperature of the measure-90° curvedment cross-sections, part in the curved respectively. pipe were T32 numbered K-type thermocouples as S3-0, S3-30, S3 were-60, usedand S3 to- measure90. The re- the hot mainingair temperature cross-sections oscillations were numbered under pulsating S1, S2, and flow. S4–S10. The outlet of the downstream pipe wasAs shownset as the in “Outlet” Figure2, section. to facilitate the subsequent discussion, the first measurement sectionTo examine named “Inlet”the inflow at theconditions outlet of of thesteady upstream and pulsating pipe is flows defined at the as the“Inlet” origin. section, S1 to S10 asrepresent shown in the Figure ten measured4, the time history cross-sections of the streamwise in the straight velocity pipe. of both The steady four sectionsflow and of the pulsating90◦ curved flow part was in measured the curved using pipe a hot were-wire numbered anemometer. as S3-0, The sampling S3-30, S3-60, frequency and S3-90. was The 10,000remaining Hz, and cross-sections the frequency were of the numbered pulsating S1, flow S2, andwas S4–S10.30 Hz. The The time outlet-averaged of the downstreamveloci- tiespipe ofwas the steady set as theand “Outlet” pulsating section. flows were almost the same. To examine the inflow conditions of steady and pulsating flows at the “Inlet” section, as shown in Figure4, the time history of the streamwise velocity of both steady flow and
Figure 4. Instantaneous velocities at the “Inlet” section of the steady flow and pulsating flow.
Energies 2021, 13, x FOR PEER REVIEW 6 of 22
Figure 3. Cross-section of pipes (from upstream view): (a,b) measurement methods; (c) measure- ment points [18].
As shown in Figure 2, to facilitate the subsequent discussion, the first measurement section named “Inlet” at the outlet of the upstream pipe is defined as the origin. S1 to S10 represent the ten measured cross-sections in the straight pipe. The four sections of the 90° curved part in the curved pipe were numbered as S3-0, S3-30, S3-60, and S3-90. The re- Energies 2021, 14, 3953 maining cross-sections were numbered S1, S2, and S4–S10. The outlet of the downstream6 of 20 pipe was set as the “Outlet” section. To examine the inflow conditions of steady and pulsating flows at the “Inlet” section, as shown in Figure 4, the time history of the streamwise velocity of both steady flow and pulsating flow was measured using a hot-wire anemometer. The sampling frequency was pulsating flow was measured using a hot-wire anemometer. The sampling frequency was 10,00010,000 Hz,Hz, and the frequency of of the the pulsating pulsating flow flow w wasas 30 30 Hz. Hz. The The time time-averaged-averaged veloci- velocities ofties the of steadythe steady and and pulsating pulsating flows flows were were almost almost the the same. same.
FigureFigure 4. InstantaneousInstantaneous velocities velocities at atthe the “Inlet” “Inlet” section section of the of thesteady steady flow flow and andpulsating pulsating flow.flow.
2.2. Data Processing Owing to the insufficient dynamic properties of the temperature sensor and the interpretation of recorded signals, which are composed of static and dynamic temperatures, the fluid temperature in pulsating flow is one of the most difficult and complex flow parameters to be measured [25]. A reliable technique studied by Tagawa et al. [26] for estimating thermocouple time constants without any dynamic calibration was applied in the present study to estimate the real hot air temperature of the pulsating flow. The first-order lag systems of the two thermocouples were used to obtain the fluid temperatures (Tg1 and Tg2) using the following equations:
T = T + τ G g1 1 1 1 , (1) Tg2 = T2 + τ2G2
where T, τ, and G represents the raw temperature, time constant, and derivative of T, respectively. The subscripts “1” and “2” represent the thermocouples of 13 µm and 25 µm in diameter, respectively. The highest similarity between Tg1 and Tg2 by maximizing the correlation coefficient R is as follows: 0 0 Tg1Tg2 R = q q (2) 0 2 0 2 Tg1 Tg2 Equation (2) was used to diminish the effect of the noise and the irrelevant signal. The time constants to maximize R should satisfy the following condition:
∂R ) ∂τ = 0 1 (3) ∂R = 0 ∂τ2
Simultaneously, the above equations can finally obtain the fluid temperatures (Tg1 and Tg2) after calculating the time constants τ1 and τ2. The time constants of the 13 and 25 µm thermocouples are 9.4 and 21.8 ms, respectively. Energies 2021, 13, x FOR PEER REVIEW 7 of 22
2.2. Data Processing Owing to the insufficient dynamic properties of the temperature sensor and the in- terpretation of recorded signals, which are composed of static and dynamic temperatures, the fluid temperature in pulsating flow is one of the most difficult and complex flow pa- rameters to be measured [25]. A reliable technique studied by Tagawa et al. [26] for esti- mating thermocouple time constants without any dynamic calibration was applied in the present study to estimate the real hot air temperature of the pulsating flow. The first-order lag systems of the two thermocouples were used to obtain the fluid temperatures (Tg1 and Tg2) using the following equations: